Bench Press Tempo: What the Evidence Actually Supports

By Szymon Spławski, a lifter and the developer of Repko, a four-phase tempo timer for iPhone. That is a commercial interest in you caring about tempo, so it is stated up here rather than buried at the bottom. Every prescription below carries a confidence rating, and the honest-gap section lists what the evidence does not support.

Start with the finding that should frame everything else: there is no randomised controlled trial of four-digit tempo prescription on the barbell bench press with strength or hypertrophy as the outcome. Not one. The 2021 Sports Medicine review on this exact topic says so in its own abstract, and nothing since has closed the gap. What does exist is a substantial body of acute work, most of it from a single laboratory, plus an unusually strong velocity-based training literature. This page separates the two and tells you which is which.

• A slower eccentric reliably lowers the weight you can lift, and nobody has published by how much. Three studies establish the direction across men and women. None of their abstracts report the effect in kilograms or percent, so "a 5-second eccentric costs you X percent" is a number that does not exist.
• Intent beats duration, and this is the strongest bench-specific evidence there is. Six weeks of matched loads and volumes, differing only in whether lifters pushed at maximal or half velocity: 18.2 percent against 9.7 percent one-repetition maximum improvement.
• Imposing an external tempo has a measured cost. In a controlled comparison, self-selected speed produced a higher one-repetition maximum than experimentally controlled speed, and 80 percent of participants rated the fully controlled condition the most uncomfortable. That is inconvenient for a page written by someone who sells a tempo timer, and it belongs here anyway.

What has actually been measured on this lift

The bench press has more acute tempo data than any other barbell lift, and almost no chronic data. Knowing which findings are which is the whole game.

A slower eccentric lowers your one-repetition maximum. Wilk and colleagues (2020) tested 90 trained men across five eccentric durations and found the maximum significantly higher with a volitional or two-second eccentric than with five, eight or ten seconds, and higher at five seconds than at eight or ten. Volitional and two seconds did not differ from each other. The same group replicated the direction in 21 strength-trained women, and Headley and colleagues (2011) found it independently in men. The direction is about as well established as anything here.

What is missing is the size of it. None of those abstracts reports the effect in kilograms or as a percentage. If you read a claim that a five-second eccentric costs you a specific percentage of your bench, that number was not taken from these studies, because these studies did not publish it.

A slow eccentric degrades the concentric that follows. At 70 percent of one-repetition maximum, mean power fell from 468 watts with a two-second eccentric to 402 watts with a six-second eccentric, and mean bar velocity from 0.60 to 0.52 metres per second. If the point of the set is to move the bar fast, a long eccentric works against it.

Tempo changes what a set actually is. Wilk and colleagues (2018) ran five sets to failure comparing 2-0-2-0, 5-0-3-0 and 6-0-4-0. Repetitions were significantly higher at the fast tempo in every set, time under tension significantly higher at both slow tempos, and the two slow tempos did not differ from each other. You cannot hold load, sets and repetitions constant and add time under tension. You are always trading.

One negative result worth keeping. Tsoukos and colleagues (2024) equated load and total time under tension between a fast and a slow conditioning set and measured the effect on a subsequent bench press throw. Electromyographic activity and applied force were significantly higher in the fast condition, and the performance enhancement afterwards was statistically the same. Higher activation during the set did not translate.

A caveat that should temper all of the above: with two exceptions, the bench-specific acute tempo literature comes from one research group, in crossover designs of ten to twenty-one participants. There is no independent replication at scale.

Bench press tempo prescriptions

Intent, not duration

The single strongest bench-specific trial in this whole area is not about tempo at all. González-Badillo and colleagues (2014) trained two groups for six weeks with identical loads and volumes, differing only in whether they pushed the concentric at maximal intended velocity or deliberately at half that speed. One-repetition maximum rose 18.2 percent in the maximal-intent group against 9.7 percent in the half-velocity group. Against heavy loads the velocity gain was 36.2 percent against 17.3 percent.

That is a very large difference for a six-week intervention, and it is about intent. It says nothing about how long the eccentric should be. It says the concentric should be pushed as hard as you can push it, which in four-digit notation is X, not a number.

This is why every prescription on this page has an X in the third position. The one part of bench tempo with strong evidence behind it is the part where you stop counting and just try to move the bar.

The pause, and the number that actually helps

Competition bench requires a pause, so the paused bench has a large practice base. A survey of 548 powerlifters found paused and slow-eccentric variations among the most commonly programmed. What it does not have is an efficacy trial: no study has compared paused against touch-and-go bench training for strength or hypertrophy over a block.

What does exist is a mechanical decay curve, and it is genuinely useful. Wilson, Elliott and Wood (1991) measured how quickly the performance benefit of the prior stretch disappears when you impose a delay at the chest. It decays as a negative exponential with a half-life of about 0.85 seconds. A one-second competition pause therefore removes roughly half the elastic contribution; a two to three second pause removes most of it.

That single number does more prescriptive work than anything else in the pause literature. If you want the paused bench to train what the paused bench is for, one second is already doing the job. Longer pauses are not more specific to competition, they are simply harder.

Two related findings are frequently misquoted. García-Ramos and colleagues (2021) found touch-and-go exceeded concentric-only one-repetition maximum by about 7.8 percent in both men and women. That is touch-and-go against a dead start from pins, not against a paused rep, and it should not be cited as the cost of pausing. Separately, Pallarés and colleagues (2014) showed a two-second stop makes one-repetition maximum testing more reliable, cutting the coefficient of variation from 4.1 to 2.9 percent. Note that a stop is not the same as a slow eccentric: Janicijevic and colleagues (2020) tested three variants and found a deliberately controlled eccentric never produced the best reliability in any comparison.

The sticking region, and what it is not

The sticking region is real, well characterised, and widely misunderstood.

van den Tillaar and Ettema (2010) timed it: it begins roughly 0.2 seconds after the bar starts moving upward and lasts about 0.9 seconds. Prime-mover activation changes measurably across the pre-sticking, sticking and post-sticking phases.

The misunderstanding is treating it as the point where you fail. The same authors (2009) compared successful attempts against failures at one-repetition maximum plus 2.5 kilograms and found the sticking region present in both, with only half of the failures actually occurring within it. Muscle activation patterns were the same in successes and failures; the differences were kinematic. The sticking region is a property of the movement, not a diagnosis of your weak point.

It also moves. Gomo and van den Tillaar (2016) found that while every grip width produced a clear sticking region, it did not occur at the same joint angles, which was the opposite of their own hypothesis. And Martínez-Cava and colleagues (2019) showed the velocity signature that defines it appears only in full range of motion, disappearing when the range is shortened.

What nobody has done is manipulate pause length or eccentric tempo and measure the sticking region as an outcome across a training block. Prescriptions built on "pause to train through your sticking point" are coach reasoning, not findings.

Velocity loss, where the bench evidence is strongest

The load-velocity relationship on the bench press is the best characterised of any lift. González-Badillo and Sánchez-Medina (2010) profiled 120 strength-trained men and found mean propulsive velocity predicted percentage of one-repetition maximum with an R-squared of 0.98, with velocity at each relative load remaining stable even after subjects got 9.3 percent stronger.

Where thresholds are concerned, the two best bench-specific trials disagree, and that is worth stating plainly. Pareja-Blanco and colleagues (2020) trained four velocity-loss groups for eight weeks and found no group difference on any dynamic strength variable, with only the 50 percent group increasing pectoralis cross-sectional area more than the zero-loss group. Rodiles-Guerrero and colleagues (2022) ran a comparable design and found a significant group difference for one-repetition maximum, largest at 25 percent velocity loss, with only the 50 percent group increasing cross-sectional area. Jukic and colleagues' 2023 meta-analysis concludes that velocity loss threshold does not influence strength gains at all, while higher thresholds modestly favour hypertrophy.

The practical read that survives all three: stop earlier if you care about strength per unit of fatigue, push closer to failure if you care about size, and do not expect the difference on strength to be large.

Two cautions. Between-participant variability in velocity at a given percentage of one-repetition maximum is markedly larger than within-participant variability, so generic velocity zones are an approximation of your own profile. And the velocity of the last repetition of a set to failure, sometimes proposed as a stopping cue, is unusable: its coefficient of variation exceeds 18 percent.

Shoulders: what the evidence does and does not say

The injury burden is real and shoulder-dominant. In competitive powerlifters, 43 percent reported problems during routine training, with shoulder, lower back and knee most affected. In Paralympic powerlifting, which is contested as bench press only and is therefore the most on-point dataset available, 61 percent of injuries were chronic overuse and 32 percent were shoulder or clavicle.

What does not exist is any evidence connecting tempo to that burden. There is no prospective study linking eccentric duration, pause length or repetition tempo to shoulder or pectoral injury on the bench press, in either direction. The pectoralis major rupture literature consists of case reports and surgical series. A 2024 narrative review on shoulder injury in the bench press recommends velocity-based and autoregulatory approaches over training to failure, and states directly that the existing work lacks robust insight into shoulder biomechanics during this lift.

So: "slow eccentrics are safer for your shoulders" is not a supported claim. Neither is its opposite. If someone tells you either with confidence, they are ahead of the evidence.

Grip width is the one adjacent variable with real data, and its effects are smaller than the internet suggests. Lehman's conclusion after measuring activation across five positions is the sensible one: given how small the changes are, grip position should be chosen to match what the athlete actually does in their sport.

The finding that argues against this page

Lee and colleagues (2023) compared self-selected speed and grip against experimentally controlled speed and grip. Self-selected produced a significantly higher one-repetition maximum. The fully controlled condition was rated most uncomfortable by 80 percent of participants, against zero percent for self-selected.

That is the closest thing in the literature to a direct test of whether imposing a tempo is a good idea, and it did not come out in favour. It measured discomfort and performance rather than adaptation, so it does not settle the question. But it is the most inconvenient result for anyone selling tempo prescription, which is exactly why it is in the body of this page rather than a footnote.

The reasonable conclusion is not that tempo prescription is worthless. It is that a prescribed tempo is a training tool with a cost, and it should be used where the cost buys something specific: competition pause work, a deliberate metabolic block, technique acquisition, or a set structure you want to keep constant across weeks. Applying a four-digit tempo to every set of every session is not supported by anything.

Common errors

Quoting a percentage for what a slow eccentric costs your bench. The direction is established. The magnitude has never been published in the abstracts of the studies that established it.

Citing the 7.8 percent touch-and-go figure as the cost of pausing. That comparison was touch-and-go against concentric-only from a dead start, not against a paused rep.

Treating the sticking region as your weak point. Half of failed maximal attempts do not fail in it, and it appears in successful attempts too.

Prescribing a slow eccentric for shoulder health. There is no outcome evidence for this in either direction.

Assuming general repetition-duration research is bench press research. The two cleanest long-duration trials on repetition duration used a machine chest press and a biceps curl.

Frequently asked questions

What tempo should I use for the bench press?

Use a controlled eccentric of about two seconds and push the concentric as hard as you can, which is 2-0-X-0. The evidence supports the explosive concentric strongly and the two-second eccentric weakly, mostly as an upper bound on how slow you can go before you start leaving weight on the bar.

Does a slow eccentric build a bigger chest?

There is no bench press trial that answers this. The general repetition-duration evidence finds hypertrophy similar anywhere from 0.5 to 8 seconds, and the two best-controlled long trials on repetition duration used a machine chest press and a biceps curl rather than a barbell bench.

How long should I pause on the chest?

The performance benefit of the prior stretch decays with a half-life of about 0.85 seconds, so a one-second pause already removes roughly half of it and a two to three second pause removes most. For competition specificity, one second does the job. No trial has compared pause durations for training adaptation.

Is the paused bench better than touch-and-go for building strength?

Nobody has tested it. Paused work is standard practice among powerlifters and it is required by competition rules, but no study has compared paused against touch-and-go bench training over a block for either strength or hypertrophy.

Will slowing my bench down protect my shoulders?

There is no evidence for that, and none against it. No prospective study links tempo, pause length or eccentric duration to shoulder or pectoral injury on this lift. The rupture literature is case reports.

The honest gap: what we don't know

There is no randomised controlled trial of tempo prescription on the barbell bench press with strength or hypertrophy as an outcome. The 2021 review of this literature says so itself, and nothing since has changed it. No trial compares paused against touch-and-go training. No trial has established an optimal pause length. Dead-stop bench press has no research presence at all. No study connects tempo to injury risk in either direction. The magnitude by which a long eccentric reduces your maximum has never been published, only its direction. Almost all of the bench-specific acute work comes from one laboratory in crossover designs of ten to twenty-one participants, without independent replication. Women appear in two of the studies cited here.

The prescriptions in the table above sit inside ranges that acute evidence and practice support. Their precise selection is coaching judgement, not a measured result.

The same standard applies to the tool. Repko has no camera, no sensor, and no way of knowing whether your second second was actually a second second. It marks each phase and trusts you to follow it, which means it removes the counting, not the discipline. A device that verifies your tempo does not exist, not in this app and not in any other. What exists is a timer that holds the number you already chose, on the rep where your form starts negotiating with you. If you were hoping for more than that, nobody can sell it to you yet.

Closing

The bench press is the lift where tempo advice is most confident and least supported. The chronic evidence does not exist. The acute evidence is real but narrow, and it mostly says that slowing the eccentric costs you load, repetitions and bar speed while adding time under tension and metabolic stress. Whether that trade produces more muscle over a training block has never been tested on this lift.

What is well supported is simpler than a four-digit prescription: push the bar hard, pause when the sport requires it, keep your test protocol consistent so the number means something, and stop the set somewhere deliberate. Use a tempo when it buys you something specific, and do not assume it is free.